Stanford, Kyle - Reference and Natural Kind Terms- The Real Essence of Locke's View

8/6/2019 Stanford, Kyle - Reference and Natural Kind Terms- The Real Essence of Locke's View

http://slidepdf.com/reader/full/stanford-kyle-reference-and-natural-kind-terms-the-real-essence-of-lockes 1/20

78

Pacific Philosophical Quarterly 79 (1998) 78–97 0031–5621/98/0100–0000

© 1998 University of Southern California and Blackwell Publishers Ltd. Published by

Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and

350 Main Street, Malden, MA 02148, USA.

Abstract: J. L. Mackie’s famous claim that Locke ‘anticipates’ Kripke’s Causal

Theory of Reference (CTR) rests, I suggest, upon a pair of important

misunderstandings. Contra Mackie, as well as the more recent accounts

of Paul Guyer and Michael Ayers, Lockean R eal Essences consist of those

features of an entity from which all of its experienceable properties canbe logically deduced; thus a substantival Real Essence consists of features

of a Real Constitution plus logically necessary objective connections

between them and features of some particular Nominal Essence.

Furthermore, what Locke actually anticipates is the most significant

contemporary challenge to the CTR: the qua-problem.

1. Locke’s ‘Anticipation’ of Kripke

Near the outset of the Third Book of An Essay Concerning Human Under-

standing, Locke draws a crucial distinction between what he calls a Real

Essence, the “unknown Constitution of Things, whereon their discoverable

Qualities depend”, and a Nominal Essence, “that abstract Idea, which the

General, or Sortal Name stands for”, that is, the set of necessary and

sufficient conditions for the application of a term (III iii 15). On this

account, the Real Essence of gold is whatever features of its constitution

are in fact responsible for its properties, while its Nominal Essence is theset of ideas of observable qualities, such as ‘heavy’, ‘yellow’, ‘malleable’,

R EFER EN CE A N D

N A T U R A L K IN D

T ER M S: T H E R EA L

ESSEN CE OF L OCK E’S

VIEW

BY

P. KYLE STANF ORD



REFERENCE AND NATURAL KIND TERMS 79

© 1998 University of Southern California and Blackwell Publishers Ltd.

and so on, which lead us to recognize a particular object as a piece of

gold. Thus, the Real Essences of substances are in the world, while their

Nominal Essences are in the mind of some observer or language-user.

In contrast to the case of simple ideas and modes, for which Real and

Nominal Essences are identical, Locke wishes to defend the controversial

thesis that the names of substances either do or should refer to objects

only by means of their Nominal Essences, and not by appeal to Real

Essences at all. As his definitions suggest, Locke’s claim is that only

Nominal Essences are available to us in grounding reference, although

we mistakenly suppose that we use our terms to pick out substances by

their Real Essences:

Fifthly, Another Abuse of Words ... We may observe, that in the general names of

Substances, whereof the nominal Essences are only known to us, when we put them into

Propositions, and affirm or deny any thing about them, we do most commonly tacitly

suppose, or intend, that they should stand for the real Essence of a certain sort of Substance.

(III x 17)

Locke quickly makes clear why this is an Abuse of Words:

the Word Man or Gold , signify nothing truly but a complex Idea of Properties, united

together in one sort of Substances: Yet there is scarce any Body in the use of these Words,but often supposes each of those names to stand for a thing having the real Essence, on

which those Properties depend. Which is so far from diminishing the Imperfection of our

Words, that by a plain Abuse, it adds to it, when we would make them stand for something,

which not being in our complex Idea, the name we use, can no ways be the sign of. (III x 18)

Locke here seems to reason that because we never come to know them,

substantival Real Essences cannot figure in the ideas we have of

substances, or in the signification of the terms we use for them.Locke’s argumentation often suggests that it is a simple empirical fact

about human beings that the references of our substance terms are not

founded upon the knowledge of substantival Real Essences. He claims,

for instance, that two samples which we name by the same substance term

can vary widely in their properties, and that this would be impossible if

we were picking out objects by their Real Essences (III vi 8). Elsewhere,

he claims that even if Real Essences were discoverable, the use of language

would considerably predate such discovery and is carried on perfectly wellby ignorant and illiterate people who have no knowledge of Real Essences.

He concludes that the use of language cannot depend upon having any

such knowledge (III vi 25). Such arguments offer the mistaken impression

that contingent epistemic limitations are Locke’s sole reason for thinking

that R eal Essences are not what ground the references of our kind terms.



80 PACIFIC PHILOSOPHICAL QUARTERLY


Indeed, when J. L. Mackie famously argues that Locke ‘anticipates’

Saul Kripke’s ‘Causal Theory of R eference’,1 he takes Locke’s arguments

against reference via Real Essences to be exclusively epistemic. He

describes Locke as correctly noting the same feature of our linguistic

practice that Kripke did in formulating the CTR, namely that we intend to annex the names of substances to their internal constitutions. And he

argues that Locke mistakenly rejects this insight, leading him to

dramatically misguided prescriptions for our use of language, solely on

the basis of an unwarranted pessimism regarding our ability to discover

substantival Real Essences.

There is a sense, of course, in which it is misleading to characterize this

as an ‘anticipation’ of the CTR: as Nicholas Jolley points out (1984, p. 151),

Locke raises the K ripkean view only to reject it wholesale, and it is Leibnizwho criticizes Locke on largely Kripkean grounds. Nevertheless, Mackie

is pushed towards crediting Locke with the anticipation of the CTR

because he is impressed with the way that Locke sets up the problem:

Mackie seems to feel that recognizing our intention to annex words to

Real Essences constitutes a giant leap in the direction of the truth about

reference (i.e., the CTR), and that Locke simply stumbles at the end, in

claiming that this intention cannot be fulfilled. It is thus in tr ibute to what

Mackie sees as a painfully near miss that he ascribes the anticipation of Kripke’s view to Locke.

In section 2 I will argue that this attribution involves a fundamental

misunderstanding of Locke’s account of Real Essences. Against Mackie,

as well as the more recent accounts of Paul Guyer and Michael Ayers, I

will argue that, for Locke, a Real Essence in general consists of those

features of an entity from which all of its observable or experienceable

properties can be logically deduced, and that a substance’s Real Essence,

therefore, consists of features of its Real Consitution plus logically necessary

objective connections between them and the features of a particular

Nominal Essence. Thus, Locke’s pessimism regarding our ability to

discover Real Essences, as he conceives them, is far from unwarranted.

In section 3 I will argue that even if we adopt the standard account of

Lockean Real Essences that I am challenging, what Locke anticipates is

in fact the most significant contemporary challenge to the CTR: the qua-

problem. He explicitly argues that our ideas of natural kinds establish thereferential significance of differences in internal constitution, and fur ther-

more, that we cannot annex a natural kind term to the unknown Real

Essence of a sample, for the efficient functioning of human language demands

that any particular natural object instantiate a variety of different natural

kinds, with distinct corresponding Real Essences.





2. M ackie, Locke and Real Real Essences

According to M ackie, Locke’s misstep resulted from an undue pessimism

regarding our ability to know Real Essences. He claims that

the general concept of stuff whose identity is given by tha t [internal] constitution is justified

by its usefulness as a framework for detailed explanations. Admittedly, such justification in

detail has come mainly since Locke’s time, by the progress of physics along lines about

which Locke, as we shall see, was pessimistic. (pp. 97–98)

Mackie seems to think, then, that if Locke had not been so pessimistic

about the prospects for our knowledge, he might well have recognized

the justification of the notion of ‘stuff whose identity is given by its internal

constitution’ and then perhaps embraced the Causal Theory of Referencewholesale. But the claim that Locke’s pessimism has proved to be misguided

misunderstands what Lockean Real Essences are supposed to be.

Mackie unequivocally equates Locke’s Real Essences with molecular

and atomic structures, and claims that Locke was simply wrong about

our capacity to discover them:

Since we can equate Locke’s real essences with what we should now call the molecular and

atomic structure of things, we may say tha t many real essences that were unknown in Locke’s

day are now pretty thoroughly known by chemists and physicists. (p. 78)

Later, Mackie repeats this conclusion in conjunction with a somewhat

harsh judgement regarding Locke’s failure to keep up with the science of

his time:

Looking back after nearly three centuries of scientific advance, we can easily see where

Locke was mistaken about these issues. Chemists and physicists have achieved the sort of

detailed knowledge of microstructure of which Locke despaired, and they have achieved itnot , in the main, by devising more powerful microscopes but by framing and testing detailed

hypotheses, a method whose power and value Locke did not realize. In fact, his philosophy

of science in this respect failed to keep up with the science of his own t ime, let alone anticipate

the future advances of physics: some of the contemporaries whom he knew and respected

were beginning to use the hypothetico-deductive method ... (p. 101)

Thus, Mackie thinks it obvious that Lockean Real Essences just are

atomic constitutions, and since we have knowledge of these, Locke’s

pessimism has turned out to be misguided. While Mackie is the mostardent proponent of this view, the unembarrassed equation or analogy

of Lockean Real Essences with modern molecular or atomic structure, or

with the genetic coding of organisms,2 is common currency in the writings

of quite a number of Locke scholars, including Woolhouse (1983, pp. 103,

113–14, 114 n. 4), Yolton (1985, p. 106), and Alexander (1985, ch. 13





(esp. pp. 277–79) and p. 296). Ironically, even Margaret Wilson (1979),

famously arguing that the role of Corpuscularian science in Locke’s

philosophy may have been emphasized to the unwarranted exclusion of

all else, both equates Real Essences with the arrangements of Boylean

corpuscles (p. 150) and claims “Now we know Locke was wrong in

thinking that the conceptual and empirical limitat ions of Boyleanism were

in principle unsurmountable” (p. 149). In each of the cited instances, the

author seems to feel that the equation of Locke’s Real Essences with

modern physical, chemical or genetic constitutions is obvious and

uncontroversial, or at least that it requires no argument. It is perhaps

unsurprising, then, that this view of Lockean Real Essences is taken for

granted in much contemporary philosophical and scientific debate. When

John Dupré attacks essentialism, for instance, he identifies Locke as the

thinker who made the distinction between Real and Nominal Essencesfamous, and claims that

Subsequent scientific history has convinced some, perhaps most, philosophers that Locke’s

skepticism was premature. Chemistry and physics have revealed a good deal about the

microstructure of things... (1993, p. 22)

I suggest that Mackie and the many philosophers who share this view

of Real Essences have misconstrued Locke’s account. Note that Locke’sexplicit definition of Real Essence is the “unknown Constitution of

Things, wheron their discoverable Qualities depend” (III iii 15). This

formulation combines two quite different ideas which Locke had far less

reason to discriminate clearly than do we. One important constituent

of Real Essence here is indeed molecular or atomic structure of a

Corpuscularian sort. But to equate Lockean Real Essences with such

molecular structures is to ignore an essential second feature: Locke has

a very strong version of the ‘dependence’ of Qualities on Real Essences

in mind. Thus he claims,

Had we ... the specifick Ideas of [substances’] real Essences in our own Minds ... to know

the Properties of Gold , it would be no more necessary, that Gold should exist, and that we

should make Experiments upon it, than it is necessary for the knowing the Properties of a

Triangle, that a Tr iangle should exist in any Matter, the Idea in our Minds would serve for

the one, as well as the other. (IV vi 11; see also IV iii 25)

This passage suggests that part of what Locke means by ‘Real Essence’is that a knowledge of it should enable us to understand and deduce its

possessor’s further properties without trial or experiment. Thus ‘discoverable

Qualities’ were supposed to ‘depend’ upon Real Essences in the strongest

possible sense: the former were supposed to be logical consequences of the

latter, and deducible from a knowledge of them in advance of experience





in just the way that the properties of geometrical figures are logically

deducible from their definitions.

Mackie views this strand in Locke’s thinking as a simple error regarding

what knowing Real Essences will get us. Because the fundamental laws

of physical interaction are synthetic and empirical, Mackie claims, Locke

was right to consider them ‘as a priori’3 as the laws of interpreted geometry

and applied mechanics, but wrong to think that the latter were a priori

at all: it was simply a mistake to think that knowledge of Real Essences

could enable us to deduce the properties of objects without trial. I would

suggest instead that this strong sense of ‘dependence’ of macroscopic

properties upon it is part of what constitutes a substantival Real Essence

for Locke: that is, that knowing the Real Essence of a thing must, by

definition, enable us to deduce its Nominally Essential properties in

advance of experience. In this case, the R eal Essences of substances wouldhave to include not only Corpuscularian constitutions, but also logically

necessary connections between those constitutions and their observable

or experienceable features, and the ‘hypothetico-deductive method’ would

allow us to achieve only a partial knowledge of Lockean Real Essences.

We find compelling textual grounds for accepting this construal of

substantival Real Essences over Mackie’s simple equation of them with

internal constitutions in Locke’s claim that the Real Essences of simple

ideas (e.g., whiteness) and modes (e.g., triangle (a simple mode), hypocrisy(a mixed mode)) are identical with their Nominal Essences (III iii 18; III

v 14).4 R. S. Woolhouse, who shares Mackie’s view of substantival Real

Essences as Corpuscularian constitutions (1971, §18), claims that the

“whole upshot of Locke’s theoretical account of mode-ideas is (whether

he recognizes it or not) that the notion of a real essence ... does not apply

to them” (p. 126), and that equating their Nominal Essences with Real

Essences therefore leads Locke ultimately into quite serious incoherence

(§23 and §26). Of course, if Locke’s Real Essences were bare Corpuscularian

constitutions, Woolhouse would be right, for modes have none.5

But Woolhouse maintains that there is no construal of Lockean Real

Essence, Corpuscularian or otherwise, which applies to modes. This is because

the problem with modes, on the view of them Woolhouse attributes to

Locke, is “not so much that their properties have no ‘tie’ or ‘foundation’

in the specifically corpuscular, mechanistic, way relevant to the properties

of substances, as that they have no ‘tie’ and ‘foundation’ at all – in this

or any analogous way” (p. 127). But Locke explicitly makes the case for

a sort of unity that he claims is possessed by modes nonetheless:it seems reasonable to enquire, whence [a mixed mode] has its Unity; and how such a precise

multitude comes to make but one Idea, since that Combination does not always exist together

in Nature. To which I answer it is plain, it has its Unity from an Act of the Mind combining

those several simple Ideas together, and considering them as one complex one, consisting

of those parts …6 (II xxii 4)





For Locke it is ‘plain’ that the unitary character of a mixed mode is

produced by an Act of the Mind. This is, of course, how we can know

the Real Essences of modes: the ‘tie’ and ‘foundation’ of their properties

is just that which we (mentally) provide. This unity is sufficient, in Locke’s

view, for the possession of a Real Essence; thus, we must ascribe to him

a notion of Real Essence broad enough to include a set of ideas of

properties unified only by a mental act.

Indeed, Woolhouse’s claim that any construal of Real Essences should

deny them to modes shows how his Corpuscularian-minded approach

gets things somewhat backwards: Locke’s claims about the Real Essences

of modes and simple ideas demonstrate the need for a unified account of

Real Essence in general which can encompass something besides mere

Corpuscularian constitutions (which modes and simple ideas do not have).

In fact, Locke’s attribution of Real Essences to modes and simple ideassuggests that it is a matter of contingent fact and not of definition that

Corpuscularian constitutions are even elements in substantival Real

Essences at all. Let us see why.

For Locke to allow that modes have Real Essences, and to equate their

Real with their Nominal Essences, his general conception of a Real

Essence must amount to something like ‘those features of a thing from

which its Nominally Essential properties can be deduced ’. Modes and simple

ideas are themselves already mental entities: notwithstanding the fact thatthey are “Dependences on, or Affections of Substances” (II xii 4), Locke

claims (perhaps mistakenly)7 that even mixed modes like gratitude, beauty

and murder are “not referred to the real Existence of Things” (III v 14),

that they have only the free combination of ideas in the mind as archetype

(IV iv 5), and that they have “their Original, and constant Existence, more

in the Thoughts of Men, than in the reality of things” (II xxii 2). Thus,

features of a mode’s (or simple idea’s) Nominal Essence are features of

the mode (or simple idea) itself , and, as Locke claims, its Real Essence,

the features of the mode or simple idea from which its Nominally Essential

features follow logically (if trivially8), will just be that very Nominal

Essence. But substances are not mental entities; therefore, features of the

Nominal Essence of a substance are not features of the substance itself.

Thus, in the case of substances, the general definition of Real Essence is

fulfilled by the conjunction of two distinct components, each of which

Locke takes to be beyond our knowledge: Corpuscularian constitutions

and the logically necessary connections between such structures and the

mental experiences we have when exposed to them – the experiences whichare collected up into the Nominal Essence of the substance. Only a

knowledge of both of these elements would enable us to deduce the

Nominally Essential features of substances in advance of experience or

experiment; thus, both must be included as components of substantival

Real Essences.





What I offer, then, is a unified, general account of Locke’s claims about

Real Essences. It seems reasonable to think that, for Locke, the Real

Essence of any sort of thing should consist of those among its features a

knowledge of which would enable us to discover, a priori, all of the ways

in which things of that sort manifest themselves in our experiences; he

does claim, after all, that Essence “may be taken for the very being of

any thing, whereby it is what it is” (III iii 15). Such a characterization

comprehends his willingness to attribute Real Essences to modes and

simple ideas, as well as his explicit claim that the Real and Nominal

Essences of such entities are identical. And on this view, while the Real

Essence of a substance consists of both a particular Real Constitution

and the necessary connections between that Constitution and some particular

set of Nominally Essential properties, it is a completely contingent fact

that it is these particular characteristics which satisfy Locke’s generalnotion of Real Essence when it is applied to substances, rather than a

matter of definition.

Several recent commentators have avoided Mackie’s simple equation

of substantival Real Essences with the internal constitutions of substances,

but have offered accounts of Lockean Real Essences which stop short of

including necessary connections between the Real Constitutions and

observable features of substances. Paul Guyer, for instance, suggests that

for Locke,

a real essence is that aspect of a particular’s real constitution which explains its possession

of those among its sensible qualities that have been singled out as comprising the nominal

essence of the species in which it is being classified by the general term by which it is

denominated. (1994, p. 133)

Michael Ayers offers a similar proposal in his (1991):

What explains the properties of the species so defined ... is corpuscularian structure ... Those

aspects of the structure of individual members of a species which they have in common and

in virtue of which they all possess the defining properties of the species, comprise what

Locke called the ‘real essence’ of the species. (vol. II, pp. 67–68)

For Guyer and Ayers, then, a substantival Real Essence is a subset of an

object’s Real Constitution: that set of features of the Real Constitution

which explains the object’s possession of the features of some particular

Nominal Essence. And on this view, Locke’s pessimism again turnsout to have been misguided, for we are indeed able to explain many

experiencable properties of objects by appeal to features of their internal

constitutions (see also Guyer, p. 134).

Of course, Locke’s claim that the Real Essences of modes and simple

ideas are identical to their Nominal Essences makes it difficult to see how





the Guyer/Ayers proposal for substantival Real Essences can countenance

a defensible account of Real Essences in general: even if we ignore the demand

that a Real Essence be selected from among features of a Corpuscularian

structure (which modes and simple ideas do not possess), it is surely

something of a strain to say that the Nominally Essential features of a

simple idea or mode ‘explain’ its possession of those very Nominally

Essential features (i.e., themselves) or that it is ‘in virtue of’ possessing

its Nominally Essential features that a simple idea or mode possesses its

Nominally Essential features. Furthermore, even if the Guyer/Ayers account

can be shown to be consistent with some defensible general account of

Lockean Real Essences,9 further textual and historical considerations

regarding Locke’s specific claims about substantival R eal Essences favor

an account of them that includes the necessary connections between

features of Real Constitutions and Nominally Essential properties.The most important such consideration is Locke’s vaunted pessimism

regarding our ability to discover the Real Essences of substances. Locke

was intimately acquainted with Corpuscularian science, and, as Mackie

himself points out above, had considerable respect for many of his

contemporaries who were beginning to fruitfully employ the hypothetico-

deductive method.10 Thus, it does not seem reasonable to suppose that

Locke was completely and unreservedly pessimistic regarding our ability

to discover the bare Corpuscularian constitut ions of objects. And indeed,when Locke discusses the state of human knowledge, he evidences a consistent

tendency to be considerably more pessimistic regarding the prospects for

discovering necessary connections between such constitutions and

observable properties than the possibility of discovering the constitutions

themselves:

Besides this ignorance of the primary Qualities of the insensible Parts of Bodies, on which

depend all their secondary Qualities, there is yet another and more incurable part of

Ignorance, which sets us more remote from a certain Knowledge of the Co-existence or

Inco-Existence (if I may so say) of different Ideas in the same Subject; and that is, that there

is no discoverable connection between any secondary Quality, and those primary Qualities

that it depends on. (IV iii 12, first italics added)11

While Locke indeed claims that we are ignorant of Corpuscularian

structures, his thoroughgoing and relentless pessimism is directed against

a ‘more incurable part of Ignorance’. This is our general inability to discover

any necessary connection between primary and secondary qualities, of whichour inability to discover such connections between the primary qualities

of the insensible parts of bodies and the observable secondary qualities

of those bodies is a salient consequence. It seems that we can make sense

of both Locke’s presumable familiarity with the advances of the hypothetico-

deductive method in discovering Corpuscularian structures and his





extreme pessimism regarding our ability to know Real Essences by pointing

out that this pessimism was directed most strongly against our ability to

learn of necessary connections between such structures and observable

properties, the crucial second component of substantival Real Essences

on the account of them I have offered.

It is important to realize that Locke’s insistence that the laws relating

primary and secondary qualities be logically necessary reflects his

acceptance of a fundamentally Rationalist view of the relations among

the elements of the natural world, a feature of Locke’s thought noted by

Woolhouse (1971, §§4 and5), Yolton (1970, pp. 79–85), Ayers (Vol II, ch.

12) and Beauchamp and R osenberg (1981, ch. 2, §III.1; ch. 3, §III), among

others. On this view, the properties of any natural object depend logically

upon it, in precisely the way that the mathematical properties of a triangle

depend logically upon its definition, and the two relations carry exactlythe same sort of logical necessity. It is for this reason that true knowledge

of a cause, for the Rationalist, permits the purely logical deduction of its

effects. This is the sort of knowledge of the natural world which God

is supposed to possess: not simply the exhaustive brute facts regarding

the existence of objects, events and properties, but a knowledge of the

objectively existing logical necessities which entail that objects, events and

properties occur as and when they do. It is this Rationalist view of the

natural world, when combined with Locke’s Empiricist awareness thatempirical science never delivers such knowledge of logical necessities,

which results in Locke’s infamous pessimism regarding the prospects for

natural science (IV xii 10): his pessimism regarding our ability to learn

the Real Essences of substances derives most fundamentally from this

general inability to learn of logically necessary connections between

features of the natural world.12

Even the claim that we have discovered features of substance’s Real

Constitutions which ‘explain’ their possession of Nominally Essential

properties obscures this Rationalist dimension of Locke’s thought: for

Locke, a Real Constitution could explain nothing without a knowledge

of the logically necessary connections between natural phenomena. Thus,

no feature(s) of any Real Constitution would suffice to explain the

possession of Nominally Essential properties, and we have not managed

to discover features which explain in Locke’s sense (proving his pessimism

to be unwarranted) so much as contemporary thinkers have largely given

up on the idea that there are any features or explanations of this kind to

be discovered.13

We should also notice that when he bemoans the ‘less incurable’ part

of our ignorance (ignorance of Corpuscularian constitutions), Locke rarely,

if ever, uses the term ‘Real Essence’. He typically uses ‘real Constitutions’,

‘minute Constitutions’, ‘the Size, Figure, Texture, and Motion of Parts’

(or some subset of these properties), or ‘the primary Qualities of insensible





Parts’. In Book IV, the term ‘Real Essence’ is typically reserved for

passages like the following:

[we could], from a Discovery of their real Essences, grasp at a time whole Sheaves; and in

bundles, comprehend the Nature and Properties of whole Species together. (IV xii 12)

Admittedly, Locke’s choice of language is notoriously slippery and at

times appears inconsistent. Nonetheless, it at least lends a blush of

plausibility to my conception of substantival Real Essences that Locke is

generally unwilling to apply this term to bare Corpuscularian constitutions,

and when he does employ it, it is in a manner consistent with conceiving

Real Essences in the way I have proposed. After all, if knowledge of Real

Essences includes the knowledge of necessary connections between

Corpuscularian constitutions and observable properties, we would expecta knowledge of them alone to enable us to ‘comprehend the Nature and

Properties of whole Species together’.

There are also specific passages in which Locke suggests that the Real

Essences of substances have the two distinct constituents I have claimed:

Had we such Ideas of Substances, as to know what real Constitutions produce those sensible

Qualities we find in them, and how those Qualities flowed from thence, we could, by the

specifick Ideas of their real Essences in our Minds, more certainly find out their Properties,

and discover what Qualities they had, or had not, than we can now by our Senses. (IV vi

11, italics added)14

Notice that Locke does use the term ‘real Essences’ here, but it seems to

encompass ‘real Constitutions’and ‘how those Qualities flowed from thence’

collectively, suggesting that both Corpuscularian constitutions and the

necessary connections between them and Qualities are constituents of this

notion.15 Only knowing both of these elements of Real Essences in the

robust sense I have proposed would, as Locke insists, enable us to deducethe Qualities of their possessors.

I would suggest that we have missed this general characterization of

Lockean Real Essences because of the very natural tendency to read our

own philosophical concerns into philosophers of the past. After all, we

have indeed achieved an impressive degree of knowledge regarding

molecular and genetic structure, while contemporary philosophical and

scientific orthodoxy has come to consider the very project of searching

for logically necessary connections between such structures and observableproperties to be nonsensical. Thus, a notion of Real Essence as

Corpuscularian molecular or genetic structure is surely the one with which

we are most concerned, and the account of substantival Real Essences

presented by Locke is near enough to be mistaken for this: Corpuscularian

constitut ions do turn out to be a constituent in Lockean substantival Real





Essences, even if this is not true by definition and even if they are not the

whole story. We have, then, mistakenly foisted upon Locke’s metaphysics

a notion of Real Essence which captures only what we ourselves have

been able to discover through advances in the natural sciences and with

which we are therefore most concerned. But as we have seen, to replace

Locke’s own robust notion of a Real Essence with a more restricted one

of greater contemporary interest does violence to his thinking and thoroughly

obscures the philosophical problems with which he was concerned.

On the account I have offered of them, Locke’s pessimism regarding

our ability to discover the Real Essences of substances appears reasonable

and defensible. But this realization leaves us with a further question of

pressing philosophical interest: if we are able to discover only bare

Corpuscularian constitutions, and not Lockean Real Essences, are such

constitutions up to the job Mackie has proposed for them, that of founding a Causal Theory of Reference? What I will now argue is that,

even if we grant Mackie his flawed sense of ‘Real Essence,’ Locke does

not anticipate the CTR, but instead offers an in principle argument against

that position which is compelling even by contemporary lights. That is,

far from anticipating the CTR, Locke refuted that position before its

birth.16

3. Anticipation or Confutation? Locke and the CTR

Mackie introduces Locke’s ‘anticipation’ of Kripke by focusing on a

mistake he thinks Locke makes regarding the difference between the

names of substances and those of mixed modes. At III vi 44–51, Locke

tells us a fable about Adam’s naming of a substance (‘zahab’) and two

mixed modes (‘kinneah’, ‘niouph’) in order to point out a difference in

the way we use these two sorts of terms. In the fable, Adam mistakenly

believes that Lamech is troubled by suspicion that his wife Adah is

committing adultery. In discussing his suspicions with Eve, Adam invents

the words ‘kinneah’ and ‘niouph’ to express the complex ideas jealousy

and adultery, respectively. In time, Adam discovers that Lamech’s trouble

has nothing to do with jealousy or adultery; nonetheless, Locke points

out, the Nominal Essences of ‘kinneah’ and ‘niouph’ continue to be those

complex ideas to which Adam first attached them.

It is quite otherwise with substances, Locke tells us. He supposes that

Adam invents the word ‘zahab’ to describe some new substance (gold)his children have brought him. Here, Locke points out , Adam is not free

to unite any set of ideas he chooses; the Nominal Essence he defines for

‘zahab’ must conform to properties actually possessed by gold. In contrast

to the cases of ‘kinneah’ and ‘niouph’, Adam must take care here that

the ideas he includes in the Nominal Essence of ‘zahab’ correspond to





actual properties of the thing he has encountered, or the Nominal Essence

he proposes will be wrong.

Locke claims that the important difference between these cases is that

for ‘zahab’, Adam has a standard imposed by nature, while he is free to

combine his ideas as he likes in naming the mixed modes. Mackie views

this as a mistake, claiming that there is a standard imposed by nature in

both cases, and that the difference is instead that, as with all substance-

terms,

He intended zahab to stand for that stuff, whatever properties and constitution it may turn

out to have; but he did not intend kinneah to stand for the sort of trouble, whatever it may

turn out to be, from which Lamech is suffering, nor niouph for whatever Adah has been up to

lately. (p. 93)

Thus Mackie claims that the real difference is in how we intend our words

to apply. He embraces the basic form of the CTR account of the naming

of natural kinds: we can say that term X will apply to this thing here and

whatever is like this thing. And as we have seen, he argues that Locke is

prevented from embracing this account only by a misguided pessimism

regarding our ability to discover the Real Essences of substances.

But this is clearly a mistake on Mackie’s part, for when Locke moves

on to consider “how and by whom these Essences come to be made” (III

vi 26), he offers more principled grounds for rejecting the CTR than mere

contingent, if incurable, ignorance. In the following passage, Locke

explicitly argues that even if we did know the Real Constitutions of

substances, this knowledge would not fix the references of natural kind

terms, for it would not settle the issue of which similarities and differences

of Real Constitution were referentially significant . Locke asks,

For what is sufficient in the inward Contrivance, to make a new Species? There are some

W atches, that are made with four Wheels, others with five: Is this a specifick difference to

the Workman? Some have Strings and Physies, and others none; some have the Balance

loose, and others regulated by a spiral Spring, and others by Hogs Bristles: Are any, or all

of these enough to make a specifick difference to the Workman, that knows each of these,

and several other different contrivances, in the internal Constitutions of W atches? ‘Tis

certain, each of these hath a real difference from the rest: But whether it be an essential, a

specifick difference or no, relates only to the complex Idea, to which the name Watch is

given: as long as they all agree in the Idea which that name stands for, and that name does

not as a generical name comprehend different Species under it, they are not essentially nor

specifically different. But if any one will make minuter Divisions from Differences, that he

knows in the internal frame of Watches; and to such precise complex Ideas, give names,

that shall prevail, they will then be new Species. (III vi 39)17

Here Locke argues that even a full knowledge of the internal constitutions

of watches would not alone allow a Watchmaker to range them into sorts.





This is because the watches all differ in internal constitution, and knowing

the Real Constitut ion of each will not tell the Watchmaker which are the

important differences between them, the differences which determine one

watch to be of a different sort than another. Particular differences of Real

Constitution can be established as referentially consequential only by

appeal to the ‘complex idea’ we annex to a particular name, even when

that idea includes features of the Real Constitution itself. As Guyer

summarizes the point of this passage, we may discover the atomic

structure of matter,

But what forces us to classify two lumps in the real constitutions of which there are the

same numbers of protons but different numbers of neutrons as two different isotopes of

the same substance rather than two different substances? Nothing but our own decision to

use the number of protons rather than neutrons as the basis of our system of classificationof the kinds of matter – a choice ... not simply forced upon us by objective similarities in

nature. (1994, p. 134, see also Ayers, Vol. II, pp. 70, 81)

Thus, Locke’s case against the CTR is not exclusively epistemic, for he

argues that even in the context of a full knowledge of the Real Constitutions

of objects and substances, the significant differences between objects, and

thus the references of our natural kind terms, can only be established by

appeal to the ‘complex idea’ or Nominal Essence to which a particular

term is annexed.

But this is not the extent of Locke’s principled argument against the

CTR, for he recognizes that the Causal Theorist’s proposal is not that

we survey the internal constitutions of objects to determine their reference-

fixing structure, but that instead we annex the reference of a particular

kind term to all those objects sharing the unknown Real Essence of a

particular sample, whatever it turns out to be. That is, the Causal

Theorist’s Adam insists that the term ‘gold’ will refer to this particular

object here and all others which share the Real Essence actually possessedby that object, whatever that Real Essence turns out to be.

But Locke also points out a fatal flaw in this fully contemporary

formulation of the central CTR strategy. Notice that in the discussion of

watches, Locke rejects any absolute division of names into specific and

generic varieties, pointing out that all that is required to change the term

‘watch’ from a specific to a generic name is for us to “make minuter

Divisions from Differences”: that is, to use some other names (with

associated Nominal Essences) to further subdivide the category of watcheson the basis of some further actual difference between them. To do so,

Locke points out, is to make ‘watch’ a ‘generical name’, but, importantly,

a name with a corresponding Real Essence and one which remains

applicable to all of the more specific sorts of watches falling under it (four-

wheeled watches, hogs’ bristles watches, etc.).





This point is developed in Locke’s explicit discussion of genera and

species. There he argues that

The same Convenience that made Men express several parcels of yellow Matter coming

from Guiny and Peru, under one name, sets them also upon making of one name, that may

comprehend both Gold, and Silver, and some other Bodies of different sorts. This is done

by leaving out those Qualities, which are peculiar to each sort; and retaining a complex

Idea, made up of those, that are common to them all ... comprehended under the name

Metal. Whereby it is plain, that Men follow not exactly the Patterns set them by Nature,

when they make their general Ideas of Substances; since there is no Body to be found, which

has barely Malleableness and Fusibility in it, without other Qualities as inseparable as those.

But Men, in making their general Ideas, seeking more the convenience of Language and

quick dispatch, by short and comprehensive signs, than the true and precise Nature of

Things, as they exist, have, in the framing of their abstract Ideas, chiefly pursued that end,

which was, to be furnished with store of general, and variously comprehensive Names. Sothat in this whole business of Genera and Species, the Genus, or more comprehensive, is but

a partial conception of what is found in the Species, and the Species, but a partial Idea of

what is to be found in each individual. (III vi 32)

Locke here sounds a favorite anti-Aristotelian theme: that the classi-

fication of natural objects is the work of humankind and fashioned to its

own purposes (‘the workmanship of the understanding’), rather than a

mirror or imposition of the natural order. But here this claim has an

important twist: Locke argues that it is inconsistent with the general

function or purpose of language for each object to fall under only a single

natural classification, despite the fact that any particular natural object

always exhibits all of its actual properties, and never merely those

represented by some general term. That is, it is the ‘convenience of

Language and quick dispatch’ which demands that we have terms referring

to a part icular object in conjunction with many different classes of others,

on the basis of different sets of properties instantiated by that object. That

is why the ‘same Convenience’ which leads us to generate a single termencompassing two separate samples of gold leads us to generate a distinct

term encompassing the various samples of metal as well. Thus, Locke

rejects any principled distinction between the generation or application

of generic and specific terms: both kinds of names for substances are

generated by abstracting away from a more complete specification of the

properties exhibited by individuals, and both reflect our common need

for terms which easily and efficiently group an individual with various

and diverse sets of others, on the basis of distinct sets of shared properties,for the purpose of collective reference.

This basic function of linguistic classification, however, is inconsistent

with the CTR’s account of reference for natural kind terms. We cannot

annex Adam’s term ‘zahab’ to the unknown Real Essence exhibited by

this particular piece of gold, because the ‘convenience of Language and





quick dispatch’ ensures that this particular sample will have many different

Real Essences: one for each general term, specific or generic, under whose

compass it falls. As Locke continues ,

If therefore any one will think, that a Man, and a Horse, and an Animal, and a Plant, etc.

are distinguished by real Essences made by Nature, he must think N ature to be very liberal

of these real Essences, making one for Body, another for an Animal, and another for a

Horse; and all these Essences liberally bestowed upon Bucephalus. But if we would rightly

consider what is done, in all these Genera and Species, or Sorts, we should find, that there

is no new Thing made, but only more or less comprehensive signs whereby we may be

enabled to express, in a few syllables, great numbers of particular Things, as they agree in

more or less general conceptions, which we have framed to that purpose. (III vi 32)18

Thus, Locke argues, it cannot be the presence of some particular unknownReal Essence which qualifies an object for the reference of a term like

‘horse’ or ‘man’. It is humankind which makes and uses languages, and

to be useful to us, the general terms of a language must be able to pick

out the very same object in many different ways: by means of various

general terms appealing to distinct sets of properties exhibited by that

object. But this fact carries a devastating consequence for the CTR, for

it implies that a sample of a substance simply has no unitary unknown

Real Essence to which the reference of a natural kind term can be annexed.Instead, we can only fix the reference of a particular natural kind term

by explicit appeal to a particular set of properties exhibited by a sample:

one of the various and diverse sets of properties by means of which we

intend to group that object with others for the purposes of collective

reference.

It is worth noting that Locke’s argument here ‘anticipates’ what is by

far the most compelling contemporary objection to the CTR: the qua-

problem. In a popular textbook on the philosophy of language, forexample, Michael Devitt and Kim Sterelny introduce and endorse the

basic reference-grounding strategy of the CTR for natural kind terms,

but then go on to point out that this strategy raises a crucial and pressing

problem:

The term is applied to the sample not only qua member of a natural kind but also qua

member of one part icular na tura l kind. Any sample of a natural kind is likely to be a sample

of many na tura l kinds; for example, the sample is not only an echidna, but also a monotreme,

a mammal, a vertebrate, and so on. In virtue of what is the grounding it in qua member of

one natural kind and not another? As a result of groundings, a term refers to all objects

having the same underlying nature as the objects in the sample. But which underlying na ture?

The samples share many. What makes the na ture responsible for the sample being an echidna

the one relevant to reference rather than the nature responsible for it being a mammal (a

nature it shares with kangaroos and elephants)? (1987, p. 73)19





This, of course, is equivalent to Locke’s worries about Bucephalus, who

is not only a horse, but also an animal and a body and a thousand other

things, and gold, which is also a metal, etc. Thus, we cannot simply annex

the reference of a natural kind term to the unknown Real Essence of a

sample, for nothing fixes the particular Real Essence of the sample which

is relevant to the grounding of that term.

Devitt and Sterelny conclude that the nature relevant to the grounding

of a particular term is established by a particular description under which

a reference-grounder “thinks of” the sample (p. 74), but they acknowledge

that this is not even close to a complete solution to the qua problem

(p. 74) and that it forces them to endorse a “descriptive-causal” rather

than a “pure-causal” account of reference (pp. 72–73). Philip Kitcher and

I have argued at length elsewhere20 that this approach, and indeed any

compelling solution to the qua-problem, vindicates much of Locke’s centralinsight concerning the reference of natural kind terms. Here, however, it

will suffice to point out that Locke’s principled grounds for rejecting the

CTR are precisely those which have forced contemporary thinkers to

reject or modify that position.

4. Conclusion

In this paper I have tried to show, in particular, that Mackie is doubly

mistaken to attribute an ‘anticipation’ of the CTR to Locke, and, in

general, that Locke’s views regarding the references of natural kind terms

are not the confused curiosities they are sometimes taken to be. I have

offered a fully general conception of Lockean Real Essences, on which

his pessimism regarding our ability to discover the Real Essences of

substances is neither misguided or uninformed, but is the natural product

of both a Rationalistic conception of the natural world and a keen

Empiricist sensitivity to the limits of empirical investigation. Perhaps more

importantly, I have argued that Locke was not regrettably thrown off the

scent of the CTR by his pessimism regarding the prospects for our

knowledge: what Locke ‘anticipates’ is not the CTR, but the most

compelling contemporary principled challenge to that view. Thus, Locke

has been damned with faint praise for conceiving of the CTR, but giving

up on it too easily: his own account of the matter is considerably more

sophisticated than that which he is credited with anticipating.

Department of Philosophy

University of California, Irvine





NOTES

I would like to thank Nicholas Jolley and Philip Kitcher for innumerable valuable

suggestions, and the members of Alan Nelson’s conference in Early Modern Philosophy

and Science (Irvine, CA; June 1997), especially Ed McCann, for an extremely helpful

discussion of this material. Any errors are, of course, my own.

1 This account of reference is usually credited both to Kripke and to Hilary Putnam(1973, 1975), but, writing in 1976, Mackie mentions only Kripke’s (1972) work.

2 Locke scholars are not alone in viewing genetic constitution as the natural candidate

for the role of a Real Essence in biological cases (see, e.g., Putnam, 1975, p. 240), but it is

worth noting that this is inconsistent with much of contemporary biological practice (see,

e.g., Mayr, 1976, 1987; Stanford, 1995).3 Here I follow Mackie in using the term a priori to characterize Locke’s views, though

this terminology is somewhat ahistorical.4 Mackie himself does not address this interpretive problem, arguing only that Locke is

wrong to think that modes have no archetypes in na ture or R eal Essences which go beyond

their Nominal Essences.5 We should note in passing that Aronson and Lewis (1970) have claimed that Locke

denies that mixed modes have Real Essences at all. This claim, however, is extremely dubious

in light of Locke’s many unqualified and unembarrassed assertions to the contrary (an

especially clear example is III v 14).6 Of course, Woolhouse may be claiming only that the unity of mixed modes is not

imposed by the physical world, but it is this very fact, along with Locke’s attribution of

Real Essences to them, which suggests that a narrow, physicalist understanding of Real

Essence is inadequate.7 Mackie and others may be right to claim that mixed modes are not, in fact, purely or

even primarily mental entities (cf. Ayers, Vol. II, ch. 8), but we must recognize that Locketakes them to be so when we try to account for his claim that their Real and Nominal

Essences (along with those of simple ideas and simple modes) are identical.8 As Locke points out, the properties of a mode may be endless, and therefore only

deducible from, rather than specified in, its Essence (e.g., II xxxii 24). This is nonetheless

consistent with the claim that a mode or simple idea’s Nominally Essential features will

follow trivially from a specification of those very features.9 See Guyer (pp. 132–33), for example, where he argues that the ‘explanation’ standard

works for simple ideas and simple modes (and that Locke is mistaken about mixed modes),

and Ayers (vol. II, pp. 57–58) where he seems to argue that identifying the Nominal and

Real Essences of modes and relations was an “entirely natural way” of identifying adistinctive role played by these Nominal Essences in enabling us to acheive instructive

conclusions by deduction alone in the purely hypothetical, a priori sciences.10 See ‘The Epistle to the Reader’ at the beginning of the Essay.11 See also IV vi 7.12 It might be thought inappropriate to use the expression ‘logically necessary’ to describe

the second constituent of Locke’s substantival Real Essences: the Rationalistic connections

between the internal constitution and the experienceable features of a substance. Nothing

hangs on this terminology, of course: what is important is that Locke took these connections

to be like the mathemat ical connections between the definition of a triangle and its various

properties (see above), that is, to exhibit a kind of necessity far stronger than the merelysynthetic, empirical or natural variety that Mackie, Guyer and Ayers suggest.

13 This suggests that the letter (although not the spirit) of the Guyer/Ayers proposal that

the Real Essence consists of those features of the Real Constitution which explain the

possession of a particular Nominal Essence could be salvaged if it abandoned the claim

that a R eal Essence was a subset of the features of a Real Constitution and recognized that

logically necessary connections between natural phenomena would be part of what





‘explained’ the possession of Nominally Essential features by substances. This revision,

however, would simply turn their proposal into mine.14 See also IV iii 16, IV iii 25, IV iv 12, and IV vi 9.15 This approach allows us to interpret passages which might otherwise suggest a Mackian

construal of Real Essences, such as the following: “we neither know the real Constitution

of the minute Parts, on which their Qualities do depend; nor, did we know them, could we

discover any necessary connexion between them, and any of the secondary Qualities” (IV iii

14). Here the first thing of which we are supposed to be ignorant is a paraphrase of the

original definition of ‘Real Essence’ (from III iii 15), which might suggest that knowledge

of Real Essences is distinct from knowledge of necessary connections. Note, however, that

Locke is unwilling to use the term ‘Real Essence’ here, using instead ‘real Constitution of

the minute Parts’. Also, notice that Locke here preserves his selective degrees of pessimism,

claiming that even if we could discover Corpuscularian constitutions, we could not discover

necessary connections between them and properties. The best understanding of Locke is to

be gotten by taking this passage as a whole to deny that we have knowledge of either of the

constituents of Real Essences.16 In what follows, I will grant Mackie the traditional conception of ‘Real Essence’ (i.e.,

equivalent to Real Constitution or features thereof) in order to avoid begging the question

against him.17 While this passage discusses an artefact (watches), Locke explicitly introduces it in ‘Of

the Names of Substances’ as simply a more familiar example for illustrating the same

problems of essential difference and species for natural kinds (substances).18 See also Ayers, Vol. II, pp. 73–74.19 Early discussions of this problem can be found in Dupré (1981) and Kitcher (1982, esp.

pp. 341–42).

20 “Refining the Causal Theory of Reference”, forthcoming in Philosophical Studies.

REFERENCES

Alexander, P. (1985). Ideas, Qualities and Corpuscles: Locke and Boyle on the ex ternal world .

Cambridge: Cambridge University Press.

Aronson, C. and Lewis, D. (1970). “Locke on mixed modes, knowledge and substances”,

Journal of the History of Philosophy, 8, pp. 193–99.

Ayers, M. (1991). Locke, 2 vols. London: Routledge.

Beauchamp, T. and Rosenberg, A. (1981). Hume and the Problem of Causation. New York:

Oxford University Press.

Devitt, M. and Sterelny, K. (1987). Language and Reality. Cambridge, Massachusetts: MIT

Press.

Dupré, J. (1981). “Natura l kinds and biological taxa”, Philosophical R eview, 90, pp. 66–90.

—— (1993). The Disorder of T hings. Cambridge, Massachusetts: Harvard University Press.

Guyer, P. (1994). “Locke’s philosophy of language”, in Vere Chappell (ed.), The Cambridge

Companion to Locke. Cambridge: Cambridge University Press, pp. 115–45.

Jolley, N. (1984). Leibniz and Locke: A study of the New Essays on Human Understanding.

Oxford: Clarendon Press.Kitcher, P. (1982). “Genes”, British Journal for the Philosophy of Science, 33, pp. 337–59.

Kripke, S. (1972/1980). Naming and Necessity (2nd ed.). Cambridge, Massachusetts: Harvard

University Press.

Locke, J. (1689/1975). An Essay Concerning Human Understanding, ed. P. H. Nidditch.

Oxford: Clarendon Press.

Mackie, J. (1976). Problems from Locke. Oxford: Clarendon Press.





Mayr, E. (1976). Evolution and the Diversity of Life: Selected Essays. Cambridge, MA:

Harvard University Press.

—— (1987). “The ontological status of species: Scientific progress and philosophical

terminology”, Biology and Philosophy, 2, pp. 145–66.

Putnam, H. (1973). “Meaning and reference”, Journal of Philosophy, 70, pp. 699–711.

—— (1975). “The meaning of ‘meaning’ ”, reprinted in Putnam (1975), M ind, Language and

Reality: Philosophical Papers, Volume 2. Cambridge: Cambridge University Press.

Stanford, P. K. (1995). “For pluralism and against realism about species”, Philosophy of

Science, 62, pp. 70–91.

Stanford, P. K. and Kitcher, P., “Refining the Causal Theory of Reference”, forthcoming

in Philosophical Studies.

Wilson, M. (1979). “Superadded properties: The limits of mechanism in Locke”, American

Philosophical Quarterly, 16, pp. 143–50.

Woolhouse, R. S. (1971). Locke’s Philosophy of Science and Knowledge. New York: Barnes

and Noble.

—— (1983). Locke. Minneapolis: University of Minnesota Press.

Yolton, J. (1970). Locke and the Compass of Human Understanding: A selective commentary

on the ‘Essay’. Cambridge: Cambridge University Press.

—— (1985). Locke: An introduction. Oxford: Basil Blackwell.

Stanford, Kyle - Reference and Natural Kind Terms- The Real Essence of Locke's View

Documents